Update C21 upper bound to 0.172422

[Dank1995]

This PR records the claimed unconditional upper bound C21 < 0.172422.

Current paper, theorem certificate, and full replay bundle (v1.1):
https://zenodo.org/records/22658444

The result is not yet independently verified, so I am submitting it with the repository's unverified (*) status.

[Dank1995]

Updated the public Zenodo record to v1.1:

https://zenodo.org/records/22658444

The claimed endpoint remains Λ < 0.172422.

Version 1.1 closes proof-provenance and reproducibility items identified during pre-release audit, including the certified B7 moment and separator Taylor-tail enclosures, restoration of the Polymath15 Eq. (20) exp(0.02y) factor in the separator error replay, explicit Proposition 6.6 reciprocal-γ provenance, and the complete 32-output direct-transition replay with interval-binding verification.

The frozen computational certificate remains byte-identical, SHA-256:

66763bb36cf35efa0cb810f145eb3a5240f2c5adfb9b3cd86d81ca0e97e866a4

The final v1.1 submission ZIP has SHA-256:

63cc29bd7cd998f736c13632712851f179c528c69c51bed2862551606651c176

I would particularly welcome scrutiny of FW-5.17, the moving separator, and the Section 5 transport.